Cpu refusing to run at high speed

[MysticDragon]

My i7-9700k refuses to run at high speeds. After a day of messing about and getting it to be fully stable at 5.1ghz, it now just cripples itself all the way down to 1.6-1.8ghz when cinebench is 50% done.

 

Anyone got any idea why? The thermals are not reaching above 90-94c.

 

Using a 750watt gold power supply that was able to handle it just fine before.

 

Edit: I have also tried to run it at stock speeds, exact same thing. And prime95 just refuses to let it run higher then 4.6.

[jordancregan193]

90-94 degrees centigrade? Yeah, that's just fine, especially when TjMax is 100 degrees. VRM temps might have something to do with it also, what board do you have? And as for Prime95, that uses AVX which probably has a multi of around 46.

[MysticDragon]

10 minutes ago, jordancregan193 said:

90-94 degrees centigrade? Yeah, that's just fine, especially when TjMax is 100 degrees. VRM temps might have something to do with it also, what board do you have? And as for Prime95, that uses AVX which probably has a multi of around 46.

z390-a

 

I know it thermal throttles to 4.6ghz when using prime95 full on cpu.

 

I am just not sure, this is not thermal throttling. It was working fine before. I dunno if it's my power supply, but then  I gotta wonder why it's acting up now 

[jordancregan193]

That isn't thermal throttling that's bringing the speed down to 4.6GHz. That's AVX, which is a much more demanding instruction set that is rarely stable at the same frequency compared to other instructions. Check your VRM temps.

[xg32]

2 minutes ago, MysticDragon said:

z390-a

 

I know it thermal throttles to 4.6ghz when using prime95 full on cpu.

 

I am just not sure, this is not thermal throttling. It was working fine before. I dunno if it's my power supply, but then  I gotta wonder why it's acting up now 

the board might be blasting too much voltage into the chip, the early z390 bios were atrocious. Also, vrm temps might be triple digits if ur cpu is 94C

[MysticDragon]

Just now, jordancregan193 said:

That isn't thermal throttling that's bringing the speed down to 4.6GHz. That's AVX, which is a much more demanding instruction set that is rarely stable at the same frequency compared to other instructions. Check your VRM temps.

It is. 1 core reaches 100c and another peaks at 100c when using prime 95

[MysticDragon]

Just now, xg32 said:

the board might be blasting too much voltage into the chip, the early z390 bios were atrocious.

The motherboard is like 1 week old, so the bios should be up to date. I also have it set at 1.36volts, something that worked fine before

[jordancregan193]

So you are hitting TjMAX. Well that's your problem then. Fix your cooling solution at stock, then wonder why it isn't working when overclocked.

[jordancregan193]

Just now, MysticDragon said:

The motherboard is like 1 week old, so the bios should be up to date. I also have it set at 1.36volts, something that worked fine before

Age of the motherboard has nothing to do with what BIOS is loaded on it. That could have been one of the first just sitting on a shelf somewhere.

[MysticDragon]

2 minutes ago, jordancregan193 said:

So you are hitting TjMAX. Well that's your problem then. Fix your cooling solution at stock, then wonder why it isn't working when overclocked.

No i'm not. I'm hitting 90-94 c on non avx, and same goes for avx so long as I am not trying out that stupid prime95 test that legit is not even healthy for your cpu cuz of the stress it unloads onto it

[jordancregan193]

You literally just said "It is. 1 core reaches 100c and another peaks at 100c when using prime 95". That's Thermal Junction Max, which is not something you should be running your CPU at. Have you figured out what your VRM temps are?

[MysticDragon]

It might have something to do with my psu? I dunno. It was drawing 180-230 watts before

[MysticDragon]

2 minutes ago, jordancregan193 said:

You literally just said "It is. 1 core reaches 100c and another peaks at 100c when using prime 95". That's Thermal Junction Max, which is not something you should be running your CPU at. Have you figured out what your VRM temps are?

When I am running the most intense test out there on the webs, it does yeah. Like 95% of all other cpus running on air. Don't exactly know how to see it

[jordancregan193]

A computer isn't stable until it's always stable. There's no such thing as "it's stable for this". I mean it could be, until it's not. Your PSU won't affect clockspeeds, if it was faulty and not providing the right amount and type of power it might bring down the system, however. The three main things are a power limit (which wouldn't bring your clocks down that far), a cpu core temp limit (which at 90-94 degrees it shouldn't be an issue) and a vrm temp limit (which you said was 50-75 degrees, so that also isn't a problem).

[MysticDragon]

Just now, jordancregan193 said:

A computer isn't stable until it's always stable. There's no such thing as "it's stable for this". I mean it could be, until it's not. Your PSU won't affect clockspeeds, if it was faulty and not providing the right amount and type of power it might bring down the system, however. The three main things are a power limit (which wouldn't bring your clocks down that far), a cpu core temp limit (which at 90-94 degrees it shouldn't be an issue) and a vrm temp limit (which you said was 50-75 degrees, so that also isn't a problem).

 

Then I honestly don't know. It was fine before, now it's not all of a sudden.  

[MysticDragon]

It seems in cinebench it just suddenly starts using 100% of the cpu half way through.

 

Is there some bullshit setting in windows that might disable it that got enabled when I activated windows? 

[MysticDragon]

Ok now the cpu is suddenly running at 100 c on 2 cores, whilst I have ran this test 30 times and got a max of 94c before. 

[MysticDragon]

I think it's my psu not being able to provide enough.

I just recalled testing out the ocing without all my 5 hard drives and 1 disk reader being installed, with is like 125 watts on it's own it no longer can take.

 

Still find it strange given my gpu (rtx 2080 ti) prob draws no where near enough when not loaded for it to matter, but hopefully buying a 1000watt+ psu will fix it

 

I'll test the OC tomorrow without my hard drives installed 

[asand1]

What CPU cooler do you have? CPUs are supposed to get hot. It's your job to cool it. If it's hitting 100c then it has plenty of power from the PSU.

[xg32]

2 hours ago, MysticDragon said:

.

ill say this once before you fry your comp, you are hitting tjmax, 94 can spike to 100 anytime. temps are 100% the problem, and running 5.1 1.36v without adaquate cooling is just insane.

[MysticDragon]

8 hours ago, asand1 said:

What CPU cooler do you have? CPUs are supposed to get hot. It's your job to cool it. If it's hitting 100c then it has plenty of power from the PSU.

 

7 hours ago, xg32 said:

ill say this once before you fry your comp, you are hitting tjmax, 94 can spike to 100 anytime. temps are 100% the problem, and running 5.1 1.36v without adaquate cooling is just insane.

I'm using the dark rock pro 4

[MysticDragon]

Honestly de-lidding and liquid metal is the only way I can see myself lowering temps now. The dark rock pro 4 is only 4 degrees hotter then the better water coolers out there, something I am not willing to touch cuz I know i'll break it

 

I'm clearly already breaking this cpu